Abstract
We studied the internal structure of rectangular frames by measuring the perceived goodness-of-fit of small probe shapes positioned within them. In Experiment 1, the rectangle's center was rated as the best-fitting position for a small, circular dot, with elevated fit ratings also along global axes of symmetry (both horizontal and vertical) and local axes of symmetry (on angle bisectors). Together, centeredness and symmetry accounted for more than 90% of the variance. Changing the probes to isosceles triangles or chevrons revealed strong orientational effects, especially when the probe's axis of symmetry aligned with the frame's axes of symmetry as mentioned above. Fit-ratings within rectangles at different orientations (0, 45 , 90 , and 135 degrees) showed that these orientational effects were driven by alignment with the rectangle's sides rather than with gravitational or retinal axes. Directional biases for better fit were also evident when the triangle pointed into the frame, upward, and/or rightward. The pattern of fit-ratings for a dot changed dramatically when a second contextual dot was also present, but high ratings were still observed when the whole configuration was symmetrical and when the two dots were close together and/or parallel to the frame's sides. When the two dots differed in size, thus breaking precise symmetry, observers gave higher ratings to positions that achieved balance in their visual weights. Alignment, proximity and symmetry were the most influential factors also when the contextual dot was replaced by a short line. When the line was extended, so that it divided the frame into two smaller rectangular parts, however, the pattern of fit ratings changed again, with the internal structure of the smaller inner rectangles dominating that of the larger outer rectangle. The implications of these results are discussed for aesthetic response as well as the perceived internal structure of shapes.